Summary
Statistical disclosure control (SDC) methods, like random rounding, are commonly used by official statistics organisations (including Stats NZ Tatauranga Aotearoa) with the goal of preventing individuals or organisations from being identified. Our project aims to develop a generic computational framework for identifying confidentiality risks in official statistics tables. Potential options to explore include: (a) a comparative analysis of the security provided by different bases for rounding (e.g., random rounding to base 5 compared to base 3); (b) comparing the risks of dependently rounded marginal totals versus independently rounded random marginals; (c) considering how random rounding compares with other SDC methods like differential privacy.
Duration and Type
- Preferable 1 semester, but can be also done in 2 semesters
- Honours project, other postgraduate project (MSc, MProfStuds, …)
Requirements
- Basic maths, statistics, CS, and machine learning skills
- Programming in R or Python
Supervisor and contact
- Daniel Wilson and Katerina Taskova
- Send CV and transcript by mail to both supervisors.